Paper Presentation 3

By : Main Mahmood Ali

SEQUENCEBIOMETRICSPALMPRINTIDENTIFICATION METHODSWHY LINE BASED APPROACHESORIENTED HAUSDORFF SIMILARITY

MEASUREFLOW DIAGRAM OF PALMPRINT

RECOGNITION SYSTEMPREPROCESSINGFEATURE EXTRACTIONMATCHINGACHIEVEMENTS

INTRODUCTIONPublic security issue

Controlled access to Personal information and data.Controlled access to work place. Like some R&D area.Prevention of cross border false infiltration.

Conventional solutions for Security implementationLock and keysPasswords and PIN codes, (Bank Account numbers, PC

passwords)Access cards, (passports, CNIC, ATM, Driving license

etc.)

Drawbacks of Traditional methodsProne to Theft/Hacking, Cloning, Forging, Forgetting Multiple passwords for different accesses for same

person.Required to be changed frequently

BIOMETRICSBiometrics is the study or a science involving

automated identification of human based on physical or behavioral characteristics.

PhysiologicalCharacteristics

Ear

Face

DNA

Finger

Iris

Palm print

TYPES OF BIOMETRICS [1]

Behavioral characteristics

Voice Pattern

Typing strokes’ Pattern

Signature

Walk or GAIT Pattern

DRAWBACKS [1]FINGER PRINT

Less area of interestHigh resolution images Effects of labor and aging

IRISUser acceptance Highly expensive hardware

FACE RECOGNITIONComplex computational situationsIllumination and OcclusionLow accuracy rate

VOICEAging and illness effectEffected by hardware / ambientnoiseLow accuracy rates

PALMPRINT [1]Larger area of interestAdditional information as

compared to fingersMore stable features than

fingersCheap Hardware than Iris

capturing devicesLow resolution imagesGreater user acceptance

PALMPRINT STRUCTURE[8]On line and off line imagesHigh and low resolution

imagesPeg based

DATABASEData base features [20]

7752 palm images 386 individuals20 samples per personTwo sub-data bases

IDENTIFICATION METHODSSub-space approaches

Feature : Wavelets, DCT, Gabor filtering etc.Subspace: PCA, LDA, ICAClassifier: Neural Networks, cosine distance,

Euclidean distanceStatistical approaches

Feature : Wavelets, Sobel filter, morphological operators etc.

Subspace: Mean and Standard deviation, Mean EnergyClassifier: Neural Networks, cosine similarity,

Euclidean distance

SCOPE OF RESEARCHEdge based palmprint recognition using

Hausdorff similarity measure

Line based approaches have additional advantages like Low resolution images are required as edges are

utilized for basic features Computationally less expensive and Better results

in illumination variation Classifier is usually a distance transform Low cost imaging devices

Classical Hausdorff distance . Gyu-Sin et al [26].

HAUSDORFF DISTANCE TRANSFORM

A B,h,B A,hmaxB A,H

baBbAa minmaxB A,h

ORIENTED HAUSDORFF SIMILARITY (OHS) . [23]

)()()(OHS .B A,H aB

Aa

aBaA ddAdA

)(OHS B A,H aBAa

daS

functionLimitingd

apositionatAimageofvectorgradientunitdA

danddAbetweenproductDotaS

aB

aA

aBaA

)(

)(

)()(

)(

FLOW DIAGRAM OF PALMPRINT RECOGNITION SYSTEM

Boundary Detection

Distance transform

Center detection

ROI extraction

ROI SEGMENTATIONROI

EXTRACTION

EDGE DETECTION

DT MAP

EDGE ORIENTATION

MATCHING

Input image

EDGE IMAGE

Roberts operators

Prewitt

Sobel

Canny[21]

CANNY EDGEProvided the edges of the

palm linesProvided the edge orientationsCatered the illumination

problem

ROI EXTRACTION

EDGE DETECTION &

EDGE ORIENTATION

DT MAP

MATCHING

DISTANCE TRANSFORM[22]Euclidian distance

Chessboard

City block/ Taxicab

221

221 )()(d yyxx

)y2– y1,– x2 x1max(d

)y2– y1– x2 x1(d

PRE PROCESSING

AND ROI

EXTRACTION

EDGE DETECTION

DT MAP

EDGE ORIENTATION

MATCHING

DISTANCE TRANSFORM

PRE PROCESSING

AND ROI

EXTRACTION

EDGE DETECTION

DT MAP

EDGE ORIENTATION

MATCHING

Distance Transform

PRE PROCESSING

AND ROI

EXTRACTION

EDGE DETECTION

DT MAP

EDGE ORIENTATION

MATCHING

MATCHING SEQUENCE ROIs of test and reference images are

extracted from input imageEdge images of test and reference images

are computed.Oriented Hausdorff similarity is calculatedThe maximum similarity is the classifier of

matchingEach 81 templates test image is matched

with of the reference image. Maximum similarity template is the

matched template

PRE PROCESSING

AND ROI

EXTRACTION

EDGE DETECTION

DT MAP

EDGE ORIENTATION

MATCHING

PALM 1 PALM 2

128x128 cropped image

128x128 cropped image

Edge Image

Distance map

INTRA CLASS LINEAR TRANSLATION

imagesROI

EXTRACTION

EDGE DETECTION

DT MAP

EDGE ORIENTATION

MATCHING

SLAVE CENTER POINTS : Linear Translation Compensation

Total of 74305 matches for imposters and 13896 for Genuine were made for recognition.

After linear translation compensation results are improved by 13%.

An EER of 2.76% has been achieved.

ACHIEVEMENTSPRE

PROCESSINGAND ROI

EXTRACTION

EDGE DETECTION

DT MAP

EDGE ORIENTATION

MATCHING

COMPARISON 

Duta et al.

Kumar et al

Zhang & Zhang

Masood et al

Proposed

Algorithm

    [11] [10] [12] [4]  

Databas

e

Subject

3 100 50 50 386

Images

30   200 500 3474

FeatureFeature points

Texture and

feature points

Wavelet signature

Texture Lines

Matching criteria

Point Pattern Matching using Euclidian distance

Euclidian distance

Euclidian distance

Wavelet (Wavelet combinat

ion)

ROHS

EER (%) 5 3 3 4.07 2.76

Q & A

REFERENCES

[1] Anil K. Jain, Patrick Flynn, Arun A. Ross, “Handbook of Biometrics,” ISBN: 978-0-387-71040-2, 2008.  [2] T. Connie, A.T.B. Jin, M.G.K. Ong, D.N.C. Ling, An automated palmprint

recognition system, Image and Vision Computing 23 (5) (2005) 501–515. [3] X.Y. Jing, D. Zhang, A face and palmprint recognition approach based on

discriminant DCT feature extraction, IEEE Transactions on Systems, Man, and Cybernetics—Part B: Cybernetics 34 (6) (2004) 2405–2415.

[4]G.M. Lu, K.Q. Wang, D. Zhang, Wavelet based independent component analysis for palmprint identification, in: Proceedings of International Conference on Machine Learning and Cybernetics, vol. 6, 2004, pp. 3547–3550.

[5] G. Feng, K. Dong, D. Hu, D. Zhang, When face are combined with palmprints: a novel biometric fusion strategy, in: Lecture Notes in Computer Science, Springer, vol. 3072, 2004, pp. 701–707.

[6] R. Chu, Z. Lei, Y. Han, S.Z. Li, Learning Gabor magnitude features for palmprint recognition, ACCV, 2007, pp. 22–31.

[7] M. Ekinci, M. Aykut, Palmprint recognition by applying wavelet subband representation and kernel PCA, Lecture Notes in Artificial Intelligence, 2007, pp. 628–642.

[8] Adams Kong, David Zhang, Mohamed Kamel, “A survey of palmprint recognition”, Pattern Recognition, Volume 43, Issue 7, 2009, pp. 1408-1418.

[9] C.C. Han, H.L. Cheng, C.L. Lin, K.C. Fan, Personal authentication using palm-print features, Pattern Recognition 36 (2) (2003) 371–381.

[10] G. Lu, K. Wang, D. Zhang, Wavelet based feature extraction for palmprint identification, in: Proceeding of Second International Conference on Image and Graphics, 2002, pp. 780–784.

[11] A. Kumar, H.C. Shen, Palmprint identification using PalmCodes, in: Proceedings of 3rd International Conference on Image and Graphics, 2004, pp. 258–261.

[12] C. Poon, D.C.M. Wong, H.C. Shen, Personal identification and verification: fusion of palmprint representations, in: Proceedings of International Conference on Biometric Authentication, 2004, pp. 782–788.

[13] J.S. Noh, K.H. Rhee, Palmprint identification algorithm using Hu invariant moments and Otsu binarization, in: Proceeding of Fourth Annual ACIS International Conference on Computer and Information Science, 2005, pp. 94–99.

[14] L. Zhang, D. Zhang, Characterization of palmprints by wavelet signatures via directional context modeling, IEEE Transactions on Systems, Man and Cybernetics, Part B 34 (3) (2004) 1335–1347.

[15] X.Wu, K.Wang, D.Zhang, “Line Feature Extraction and Matching in Palmprint”, in: Proceeding of the Second International Conference on Image and Graphics, 2002, pp.583–590.

[16] Laura Liu and David Zhang “Palm-Line Detection”, International Conference on Image Processing (ICIP) : Genova, Italy, v. 3, 2005, pp. 269-272.

[17] Fang Li, Maylor K.H. Leung, Xiaozhou Yu,” Palmprint Matching Using Line Features” The 8th International Conference ,Advanced Communication Technology, ICACT 2006.

[18] David Zhang, Wai-Kin Kong, Jane You, and Michael Wong, “Online Palmprint Identification”, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 25, no. 9, September 2003.pp1041-1050.

[19]. Fang Li, Maylor K.H. Leung, “Two-stage Approach for Palm print Identification” ICARCV 9th International Conference on Control, Automation, Robotics and Vision, December 2006.

[20] PolyU Palmprint database, presently available at:http://www.comp.poly.edu.hk/biometrics/

[21] J. F. Canny. “A Computational Approach to Edge Detection”. IEEE. Transaction. Pattern Analysis and Machine Intelligence, 1986, pp. 679-698,.

[22] Rafael C. Gonzalez, Richard E. Woods, Steven L. Eddins. “Digital Image Processing Using MATLAB”, Second Edition 2007, printed by Dorling Kindersley (India) Pvt. Ltd. Under license from Pearson Education, Inc.

[23] Dong-Gyu-Sin, Rae-Hong Park “Oriented Hausdorff Similarity Measure of object matching”. MVA ’98 IAPR workshop on Machine Vision Applications, Makuhari, Chiba, Japan, 1998.

[24] Ajay Kumar, David C. M. Wong, Helen C. Shen, and Anil K. Jain, “Personal Verification Using Palmprint and Hand Geometry Biometric, Proc. 4th Intl. Conf. Audio- and Video-Based Biometric Authentication (AVBPA), Guildford, UK, 2003 , pp. 668-675.

[25] Zhang, and D. Zhang, “Characterization Of Palmprint By Wavelet Signatures Via Directional Context Modeling”, IEEE TSMC(B), 34(3), 2004. pp.1335-1347

[26] H Masood, M Mumtaz, M A Afzal Butt, AB Mansoor and S A Khan, “ Wavelet Based Palmprint Authentication System”. Proc. of IEEE. International Symposium of Biometric and security Technologies, Islamabad, Pakistan, 2008.

[27] M A Asif and H Masood ,“Palmprint Identification Using Contourlet Transform” 2nd IEEE International Conference on Biometrics Theory, Applications and Systems, BTAS. 2008.pp.1-5.